Association for Public Policy Analysis & Management

In the United States, non-English speakers are an important part of the overall population. Thirteen percent of the total U.S. population in 2012 were foreign born. Of those, 29% (about 12 million) did not speak English well or at all (American Community Survey, 2012). Because of the over-representation of limited English households in many federal surveys and program evaluations, the ability to include these households can be crucial to the success of a study.

Policy researchers are often faced with the need to collect data from linguistically heterogeneous populations and populations that do not speak the same language as the researchers. Despite this, they must assure full data comparability in their studies regardless of the language of data collection. Since data collection tools are crucial components of an evaluation research design the research designer has a central role in making sure data collection tools are translated following high quality methods that assure data comparability. 

Despite the attention that is given to well thought out and designed data collection instruments, translation is often regarded as a step that comes after instrument design, and gets contracted out without much control or methodological thinking. However, translation and/or adaptation of survey questionnaires are central steps of instrument development that require planning and careful execution. This presentation will discuss how to bring rigor and systematization to the translation process and what are best practices for research instrument translation and translation assessment.